The invention relates to a receiver for digital signals which have successively undergone, on transmission, a channel coding and a differential coding and which resulting coded digital signals have been assigned to symbols of a constellation, the receiver comprising differential decoding means which calculate metrics in response to received symbols, and channel decoding means for producing decoded digital signals in response to the calculated metrics.
Communication techniques for transmission through the air or by cable use a modulation of one or various carriers by means of digitally coded symbols. Among these techniques one preferably uses a differential modulation for which the phase of a transmit symbol is coded relative to the phase of the symbol prior to transmission. This presents the advantage to decode received symbols on reception by determining phase differences between the symbols. The binary data to be transmitted generally undergo various processes intended to ensure a transmission reliability under generally disturbed transmitting conditions. Among them a convolution coding is often carried out followed by a time-division/frequency-division interleaving of the symbols. Alternatively, a block coding may be carried out.
Such a process is described in the document FR-A-2 639 495 in the case of an application to a multicarrier transmission. Considering in this document only the aspect relating to this invention, there is indicated that at the receiver end a differential demodulation takes place of received symbols which is followed by a time-division/frequency-division deinterleaving of the symbols and a maximum likelihood decoding called Viterbi decoding.
The differential demodulation described consists of generating, in response to a series of received symbols r.sub.0,r.sub.1 . . . r.sub.N, a series r'.sub.1 . . . r'.sub.N, so that r.sub.k '=r.sub.k r.sub.k-1 * where r.sub.k-1 * designates the complex conjugate of r.sub.k-1 and k designates a current index.
For each received symbol r.sub.k, based on the series r'.sub.1 . . . r'.sub.N, a real part Re(r'.sub.k) is determined which is a metric with which is associated a coded information bit and also an imaginary part Im(r'.sub.k) is determined which is a metric with which another coded information bit is associated. The metric sequences thus formed thereafter undergo a deinterleaving followed by a convolution decoding which restores the bits of initial information. In the case of convolution coding a Viterbi decoder is used.
But these metrics are approximated metrics which provides that the decoding does not produce an optimum result.